[PATCH] drm/amdgpu: do not disable SMU on vm reboot

[Nirmoy]

Tested Device passthrough with polaris(rx 570), the patch had no effect.

With or without the patch I could only do passthrough 3 times before the 
host throws

IOMMU error.


[ 210.762150] vfio-pci 0000:01:00.1: can't change power state from D0 to 
D3hot (config space inaccessible)

[ 210.762732] vfio-pci 0000:01:00.0: can't change power state from D0 to 
D3hot (config space inaccessible)

[ 210.899735] AMD-Vi: Completion-Wait loop timed out

[ 211.022824] AMD-Vi: Completion-Wait loop timed out

[ 211.145434] AMD-Vi: Completion-Wait loop timed out

[ 211.268250] AMD-Vi: Completion-Wait loop timed out

[ 211.778593] iommu ivhd0: AMD-Vi: Event logged [IOTLB_INV_TIMEOUT 
device=01:00.0 address=0x107d5fb50]

[ 212.780494] iommu ivhd0: AMD-Vi: Event logged [IOTLB_INV_TIMEOUT 
device=01:00.0 address=0x107d5fb90]

[ 212.781060] iommu ivhd0: AMD-Vi: Event logged [IOTLB_INV_TIMEOUT 
device=01:00.0 address=0x107d5fbb0]


Regards,

Nirmoy

On 7/23/20 4:22 PM, Nirmoy wrote:
> Hi Alex,
>
> On 7/23/20 3:59 PM, Alex Deucher wrote:
>> Can you check and make sure this doesn't break polaris or some other 
>> generation?
>
>
> Do you mean breaking device passthrough or normal operation ? I have a 
> rx 570,
>
> I can test on that.
>
>
> Regards,
>
> Nirmoy
>
>
>>
>> Alex
>>
>> On Wed, Jul 15, 2020 at 1:12 AM Quan, Evan <Evan.Quan at amd.com> wrote:
>>> [AMD Official Use Only - Internal Distribution Only]
>>>
>>> Acked-by: Evan Quan <evan.quan at amd.com>
>>>
>>> -----Original Message-----
>>> From: Nirmoy Das <nirmoy.aiemd at gmail.com>
>>> Sent: Friday, July 10, 2020 10:51 PM
>>> To: amd-gfx at lists.freedesktop.org
>>> Cc: Deucher, Alexander <Alexander.Deucher at amd.com>; Quan, Evan 
>>> <Evan.Quan at amd.com>; Das, Nirmoy <Nirmoy.Das at amd.com>
>>> Subject: [PATCH] drm/amdgpu: do not disable SMU on vm reboot
>>>
>>> For passthrough device,  we do baco reset after 1st vm boot so
>>> if we disable SMU on 1st VM shutdown baco reset will fail for
>>> 2nd vm boot.
>>>
>>> Signed-off-by: Nirmoy Das <nirmoy.das at amd.com>
>>> ---
>>>   drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
>>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c 
>>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> index d1adbc45d37b..07be61ce969a 100644
>>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>>> @@ -1178,7 +1178,8 @@ amdgpu_pci_shutdown(struct pci_dev *pdev)
>>>    * unfortunately we can't detect certain
>>>    * hypervisors so just do this all the time.
>>>    */
>>> -adev->mp1_state = PP_MP1_STATE_UNLOAD;
>>> +if (!amdgpu_passthrough(adev))
>>> +adev->mp1_state = PP_MP1_STATE_UNLOAD;
>>>   amdgpu_device_ip_suspend(adev);
>>>   adev->mp1_state = PP_MP1_STATE_NONE;
>>>   }
>>> -- 
>>> 2.27.0
>>>
>>> _______________________________________________
>>> amd-gfx mailing list
>>> amd-gfx at lists.freedesktop.org
>>> https://nam11.safelinks.protection.outlook.com/?url=https%3A%2F%2Flists.freedesktop.org%2Fmailman%2Flistinfo%2Famd-gfx&data=02%7C01%7CNirmoy.Das%40amd.com%7Cf4018ca31bdf481e0bba08d82f1096b8%7C3dd8961fe4884e608e11a82d994e183d%7C0%7C0%7C637311095591342103&sdata=PvRVEMzHPeQnQC%2Bf7DI1Y3vnEZLnp9Af%2F07KAIQjKsM%3D&reserved=0 
>>>